Potent antiretroviral medications provide remarkable potential benefit to HIV-infected persons in terms of decreased morbidity and mortality. However, adherence is critical to reaping the fruits of these treatments. Nonadherence results in the development of resistant HIV strains and clinical failure. In Harlem the barriers to adherence are significant and include: substance use, fragile social supports, distrust of the health care system and difficulty in communicating with providers. In addition the complexity of these regimens i.e. number of pills and associated strict instructions are an added challenge. We propose to utilize the Transtheoretical Model of Change (TMC) as the basis for an experimental intervention to promote adherence to antiretroviral medications. This model was developed to model stages of commitment and recommitment to substance use therapy and is compatible with periods of adherence/nonadherence/discontinuation/resumption noted with antiretroviral therapy. We will enrich this model with stronger helping relation and social support elements. Our study will use a randomized clinical trial to compare an experimental intervention applying the TMC model to current clinical practice. The experimental intervention consists of two main components: social support and assistance in overcoming barriers. Based on the baseline stage of behavior change, specific elements of the experimental intervention will be provided by study caseworkers, support groups and peers workers. Participants will be HIV-infected persons who are eligible for antiretroviral therapy. One hundred and sixty participants will be randomized, stratified by their TMC stage of behavior change, to either the experimental group or current clinical practice for 12 months duration. Follow-up will continue for an additional 6 months to assess the durability of the intervention effect. The primary aim of the study will be to assess the effectiveness of the intervention on adherence with antiretroviral medications as defined by achieving at least 90 percent adherence with medication doses. In addition, we will identify participant characteristics (including substance use) that are associated with adherence and the impact of the intervention on outcomes, particularly substance use. The primary outcome will be assessed via a self-reported adherence questionnaire. Assessment of adherence will also be measured by confidential participant survey, as well as by provider and peer assessments. Biological surrogates of adherence will include plasma viral load and HIV genotypic resistance patterns. Additional questionnaires will be used to evaluate participant substance use, quality of life, mental health/depression, health care utilization, social support networks, participant satisfaction and self efficacy. This study provides an opportunity to assess a practical intervention, based on a theoretical framework, for maximizing adherence in a minority population with high substance use rates, a group currently bearing the brunt of the epidemic.